+++ This bug was initially created as a clone of Bug #1345730 +++ The most frequent failure modes in bug 1345730 were addressed, but some other, less-frequent crashes continue.

Talos crash frequency seemed to increase again around April 28. Most crashes are on Windows. There is some variation in the crash reports, but the majority are timeouts in Windows 7 e10s g2; these crash reports have WinUtils::WaitForMessage() in them, which seems reasonable -- probably just waiting for a message when the test times out and the browser is killed. Examples: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-central&job_id=95613034&lineNumber=7437 https://treeherder.mozilla.org/logviewer.html#?repo=autoland&job_id=95362228&lineNumber=2426 https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=95354796&lineNumber=2066 https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=95100900&lineNumber=2438 https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=95071366&lineNumber=2153

It looks like those g2 timeouts started around https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=88c2a0b2dde53837f770d76050f35aaa18c03bbd&filter-searchStr=windows+talos+e10s. Bug 1319111 was backed out for this, but failures continued.

And now I see :philor figured this out and backed out bug 1337062. So, false alarm, all those g2 timeouts should stop now.

Recent failures here are typically win7 e10s g2 and crashes are only due to timeouts. Crash reports are available, but look unremarkable, to me.